Well, given that there's no need to do web through CGI anymore, I didn't think that web would be the proper use case to compare. Comparing CGI shell scripts to ASP.NET is really disingenuous.

A wonderful example of how applications today get written in completely unsuitable languages. (Why not write it in something that can keep the non-derivable part of the application's state without explicit checkpoints woven throughout the whole thing?)

It has reincarnated as DynaBook Jr. at VPRI. But you'll have to wait for it.

So are programming language implementations.

That would improve the situation somewhat, but it still won't help you much if the target has an high-Isp, high-thrust propulsion system. Those are difficult to scale down. A nuclear reactor, for example, can't be manufactured below a certain size, so your missile would have to be ship-sized (and very expensive). That gets you back to trying to cover most of the distance with a high-speed coasting projectile with terminal guidance. This would also have the benefit of the projectile being mostly inert, therefore difficult to detect at longer distances, but you still have the "what if the target starts accelerating shortly after I fire" problem. Perhaps shooting multiple projectiles to account for the possible target's state space evolution would be an answer - after all, a single precise hit could be all it takes.

To those people, every point seems to matter. They'll apparently do absolutely anything...because everyone else does it too! (Well, that's free market to you and me.)

But not in cash. What about counting it as health insurance payments? Let's call it barter.

You get two iPoints.

http://en.wikipedia.org/wiki/R... Given a fixed solid-fuel engine thrust, simply perform guidance by attitude change. You don't need thrust vectoring for that purpose. It might not even be accurate enough - when vectoring the solid-fuel engine, the moment you'd get for attitude control would be a function of not just the nozzle angle but also of the oscillating thrust. Dealing with oscillating acceleration would be tricky enough with precise attitude control.

And you used a time machine to change that?

Oh no, controlling the trajectory by altering the thrust direction most certainly works, and you don't need thrust vectoring for that, RCS is enough. The problem is that with solid fuel chemical engines, once you start them, you're committed. Basically it means that you can only engage them within a fairly small envelope around the target during the final approach: you can't use them to correct the trajectory of the long coasting phase which for long-range engagements would probably form the vast majority of the weapon's range. If the target engages a high-energy drive during the coasting phase, there's a chance (I'd have to figure out some rough estimates of sensitivity for that) that you'd miss for being unable to follow suit. It's an interesting problem, though - definitely deserves some further study.

I wonder how many Duff's devices are in it. ;-)

And it comes for a price: PowerShell isn't exactly a speed demon.

Have you seen scsh?

The problem with bash is that even more than most shells (perhaps except for zsh), it's exceedingly obscure and baroque. This kind of featuritis, combined with the N^2 interaction between the individual features, tends to lead to unexpected consequences. What happened to "keep things simple" with shells?